Evaluation of non-stationary spatial relationship between meteorological-environmental parameters and PM2.5

Abstract

Particulate Matter with diameters less than 2.5 µm (PM2.5) is a harmful pollutant sensitive to environmental and meteorological changes. Understanding how these parameters correlate with PM2.5 levels is crucial for effective air quality (AQ) monitoring. This study employs a methodology to assess both seasonal variations as well as spatial non-stationarity of associations between meteorological (maximum and minimum air temperatures, precipitation, surface pressure, u- and v-component of wind) and environmental parameters (Normalized Difference Vegetation Index and Land Surface Temperature) with PM2.5 concentrations. This study was conducted in the densely populated north Indian states of Delhi, Haryana, and Uttar Pradesh. The satellite-based input data of this region was accessed and utilized using cloud- based computational platform of Google Earth Engine (GEE). The methodology initially involved the utilization of an Ordinary Least Square (OLS) model to identify significant parameters followed by the implementation of a Geographically Weighted Regression (GWR) to assess the spatial non-stationarity of the relationship of PM2.5 concentrations with environmental and meteorological parameters. During both the summer and winter seasons, strong correlations were observed between PM2.5 concentrations and the OLS identified significant parameters (R2summer = 0.93, R2winter = 0.94). Also, the mean PM2.5 concentration during the period of 2019–2020 was 73.98 µg/m3 in summer and 154.04 µg/m3 in winter. The study also highlighted the impact of anthropogenic influences on PM2.5 concentrations, as evidenced by a decreased concentrations during the COVID-19 lockdown period. Additionally, hotspot analysis was performed to identify 23 sub-divisions (Taluks) as prioritized zones with low AQ based on PM2.5 concentrations. These findings aid policymakers and planners in implementing effective mitigation measures and offer insights applicable to other cities' air quality assessments.